Ask LLOG: "Big dumb hat" v. "Dumb little dog"

« previous post | next post »

From T.S.:

I have read before about English’s very rigid adjective order – we say “nice green chair” not “green nice chair”.

A recent (not very funny) sketch on Saturday Night Live featured Amy Schumer extolling the virtues of wearing a “Big dumb hat”. The punchline was that this accessorises perfectly with a “Dumb little dog”.

“Big dumb hat” sounds right and “Dumb big hat” sound wrong.

“Dumb little dog” sounds right and “Little dumb dog” sounds wrong.

Whither English’s rigid adjective order?

Short answer: It's complicated.

Longer answer: There's a long history of work on related topics, starting with Pāṇini. A classic and accessible reference is Cooper and Ross's brilliant 1975 paper "World Order". They focus primarily on the order of elements in conjunctions ("bigger and better", "fore and aft", "cat and mouse", "now and then", "here and there", …), and describe a complex web of semantic and phonological influences. They cite Jespersen in support of the idea that such "freezing" also applies "in compound words, particularly compounds involving reduplication" ("namby-pamby", "razzle-dazzle", "hickory-dickory-dock"). And they also note (pp. 94-96) that "It seems safe to conclude […] that at least some of the principles governing the ordering of conjuncts and the ordering of prenominal adjectives are the same."

On the phonological side, they list the following constraints (p.71):

Compared to place 1 elements, place 2 elements contain, other factors being equal:

  1.   more syllables (Pāṇini's law)
  2.   longer resonant nuclei
  3.   more initial consonants
  4.   a more obstruent initial element, if both place 1 and place 2 elements start with only one consonant
  5.    vowel containing a lower second formant frequency
  6.   fewer final consonants
  7.   a less obstruent final segment, if both place 1 and place 2 elements end in a single consonant

Applying these principles to the cited pair of examples, we see that:

In combining big and dumb, principles (5) and (7)  apply, so we prefer "big dumb". (And maybe principle (2) as well…)

In combining little and dumb, principle (1) applies, so we prefer "dumb little".

There are some semantic principles that also apply, from the 1975 paper and from other work since then — but I'll leave it there for now, except to add that the cited "nice green" vs. "?green nice" preference might be due to the subjectivity constraint, as well as principles (3) and (7) from the phonological list above.

Those who want more can consult some of the 500-odd works that cite "World Order". I was especially struck by Kwon and Matsuda's 2019 paper "On the ordering of elements in ideophonic echo-words versus prosaic dvandva compounds, with special reference to Korean and Japanese". The abstract:

Building on Childs’s (Pragmat Soc 5(3):341–354, 2014) proposal that skewed phonotactic distributions provide a legitimate resource for expressiveness in ideophones, often described as iconic words, this study examines whether there are differences in element ordering between ideophonic echo-words and prosaic dvandva compounds, with special reference to Korean and Japanese. Measured against Cooper and Ross’s (in: Papers from the parasession on functionalism, Chicago Linguistic Society, Chicago, pp 63–111, 1975) claimed-to-be-universal phonological constraints for the ordering of conjoined elements pertaining to element-initial consonants and vowels, the study reveals that both Korean and Japanese data comply with the constraints in general. However, in Korean, echo-words are significantly different from dvandva compounds in their compliance with the consonant constraint while they are not so with the vowel constraint. In reverse, echowords and dvandva compounds in Japanese show a significant difference in their compliance with the vowel constraint but not with the consonant constraint. The findings provide quantitative evidence for the cross-linguistic applicability of the proposed phonological principles for element ordering and the language-specific phonotactic deviance of ideophones vis-a`-vis the matrix language for the preferred ordering patterns.



  1. Linda Death said,

    November 14, 2022 @ 9:34 am

    Does dumb mean the same in both sentences? I am in the UK where dumb still means unable to speak, though I believe that in some places it has shifted to mean stupid.

    I can see that a hat and a dog can both be unable to speak, but a hat wouldn't be stupid. So what others meanings am I missing?

  2. J.W. Brewer said,

    November 14, 2022 @ 10:53 am

    Re the subjectivity constraint, it strikes me that "dumb" in both examples is more subjective than big/little, so the tendency of the less subjective adjective to be closer to the noun modified is apparently not strong enough to overcome these phonological factors in "big dumb hat." Likewise any other semantically-based account of adjective order doesn't explain the difference here and thus does not reflect a tendency sufficiently strong to overcome the other factors mentioned.

  3. Terry K. said,

    November 14, 2022 @ 12:07 pm

    @Linda Death

    I think that's a good question, whether dumb means the same thing in both sentences.

    I think it's definitely in the "stupid" realm, rather than the "can't speak" realm for both, but I think the meaning is a little different. A dog can be stupid/dumb in the lacking-brainpower sense. Versus calling a hat dumb or stupid is more figurative.

    Also, though a hat can't speak it would be very odd to call a hat "dumb" in that sense, since hat's don't speak. (Same for dogs, though it's possible some might use "dumb" to mean can't bark. Not in American English, though, where a dumb dog is a dog lacking in thinking ability.)

  4. Sergey said,

    November 14, 2022 @ 12:17 pm

    I'm not a native English speaker, but to me it feels like the characteristics that are more closely connected to the noun go closer to it, and those more related to the "whole package" of already partially characterized noun go farther.

    So "big [dumb hat]" feels to me like that the hat is dumb, irrelevant of the size, and it also happens to be big. But "dumb [little dog]" feels that the dumb thing is not a dog as such but specifically a little dog, because of specific qualities of the little dogs. If a hat were dumb because of its size being inappropriate to a situation (think of a sombrero on a packed bus), "dumb [big hat]" would probably be more suitable.

  5. James H. said,

    November 14, 2022 @ 12:24 pm

    That hat makes you look stupid, so it looks dumb on you. That makes it a dumb looking hat. Or just a dumb hat, as the "looking" doesn't seem to add anything to the meaning. You could say "stupid hat" just as easily. "Silly hat" works, but would mean something more like one of those beanies with a little propeller on top, whereas "dumb hat" would mean one that doesn't suit you.

  6. cervantes said,

    November 14, 2022 @ 12:27 pm

    It's not that the hat itself is dumb — i.e. stupid — it's that it makes the person wearing it seem dumb. Same with the dog, the adjective doesn't apply to the animal but to its owner.

    In both of these cases, there is a semantic issue. Little dumb dog could indeed mean that the dog is stupid. If we've already talked about your big hat in the past, I would likely say "Why are you wearing that dumb big hat again?"

  7. Mark Liberman said,

    November 14, 2022 @ 1:33 pm

    @cervantes: A good example of how the adjective-order constraints are not really absolute, but can be (and often are) changed by information-structure issues. For more, see Liberman & Sproat 1991, e.g.

  8. Philip Taylor said,

    November 14, 2022 @ 1:55 pm

    Linda — my understanding is that American "dumb" is derived from German "dumm" =[ stupid], not English "dumb" [= unable to speak], whence the difference in meaning.

  9. Rick Rubenstein said,

    November 14, 2022 @ 4:17 pm

    Making this example particularly difficult to account for semantically: To my ear, "dumb little dog" is indeed the natural order, but "big dumb dog" sounds better than "dumb big dog". I think phonetics wins on that one.

  10. Jerry Packard said,

    November 14, 2022 @ 6:34 pm

    I think the two meanings of ‘dumb’ here are ‘stupid’ and ‘undesirable.’ As for ‘dumb little’ vs. ‘little dumb’ , I think that the syllable constraint takes care of the lion’s share of the variance.

  11. Jerry Packard said,

    November 14, 2022 @ 6:54 pm


    ‘If we've already talked about your big hat in the past, I would likely say "Why are you wearing that dumb big hat again?" ‘

    ‘dumb’ becomes old information, and so it is more acceptable to occur on the left.

  12. J.W. Brewer said,

    November 14, 2022 @ 7:23 pm

    FWIW, the google n-gram viewer indicates that the bigram "big dumb" has been substantially more common than "dumb big" since at least the mid 19th century. Someone with more spare time than I have at present could drill down into the hits for both and try to see if there's any coherent pattern regarding when the "dumb big" variant is or isn't used. The viewer claims there aren't any hits in the corpus where either combo is immediately followed by "hat."

  13. J.W. Brewer said,

    November 14, 2022 @ 7:38 pm

    Should have added that also per the google n-gram viewer (with all its limitations), "little dumb" was consistently more common than "dumb little" from 1808 through 1955, then there was a disputed lead-swapping interlude until 1968, following which "dumb little" pulled into the lead and stayed in the lead. So maybe these patterns and tendencies (and which factor outweighs which when they point in different directions) are not static and timeless?

    [(myl) One of the "limitations" of the n-gram viewer in this case: "little dumb" is often a substring of "a little dumb", where "a little" is an adverbial modifier of "dumb". As in this song. In other irrelevant hits, "a little" can be a quantifier for continuations like "dumb luck"; or "little" can be an adjectival modifier of collocations and fixed phrases like "dumb dumb", "dumb ass", "dumb question", etc.]

  14. Jerry Packard said,

    November 14, 2022 @ 8:20 pm


  15. Jim said,

    November 14, 2022 @ 8:26 pm

    I have never analyzed it, but this adjective order is clearly why I despise the George Strait song "Blue Clear Sky", a presumably intentional disruption of "clear blue sky".

    The inversion is very memorable, but only because it annoys the snot out of me. And I can handle fingernails on a chalkboard.

  16. Jonathan Smith said,

    November 14, 2022 @ 11:06 pm

    One issue may be that the assumption that Adj1 Adj2 N simply denotes an intersection i.e. a N that is both Adj1 and Adj2 is not what these arrangements generally (ever?) actually mean…

    To take "nice green chair," a possible application is "That's a nice green chair", in which it seems "green chair" is to establish joint attention and the chair being "nice" is the real predicative content — so ”hey, that green chair, it's nice."

    A learner of Chinese quickly finds that you can't stack up adjectival modifiers in the manner you might expect given English, such that literally "That's a nice green chair" (say, nà shì yī zhāng bù cuò de lǜ dèngzi) is not wrong but certainly clumsy and weird — one should of course simply say nà zhāng lǜ dèngzi bù cuò lit. "that green chair, [it's] nice" — go figure :D

  17. Andrew Usher said,

    November 15, 2022 @ 8:16 am

    Certainly, Sergey is on to something when he mentions the different shades of meaning 'dumb' can carry. But in this case I think one would say 'dumb little' even if the dog's actual intelligence (of which a hat has none) were being referred to.

    The sound pattern also matters, per Jerry Packard. 'Big; resists being destressed, while English very much likes to put the stress on the first adjective. 'Little' doesn't mind.

    Also to Philip Taylor:
    It may be that dumb meaning stupid is wholly a borrowing from German dumm but that simply transfers the required sense evolution there, as they are definitely cognate and 'mute' presumably the original meaning.

    k_over_hbarc at

  18. Francois Lang said,

    November 15, 2022 @ 8:16 am


  19. Philip Anderson said,

    November 16, 2022 @ 8:18 am

    Although dumb meaning stupid is slang, it is certainly used in the UK with this meaning, as well as being a synonym for mute.
    You can’t get more British (or conservative) than Private Eye, yet it has a Dumb Britain section:

  20. Bloix said,

    November 16, 2022 @ 6:14 pm

    "Dumb" can mean no more than "annoying," in the same way that "fucking" means annoying. It's a bit childish, perhaps.

  21. Philip Taylor said,

    November 17, 2022 @ 5:44 am

    Can "fucking" mean "annoying", Bloix (and if so, can you give an example) ? I ask because, trying to think of a usage in which it did mean "annoying", I came up with only the following, in which it means "annoyingly" (advb.) not "annoying" (adj.) —

    "Oh, that is fucking stupid !"

  22. Philip Anderson said,

    November 17, 2022 @ 8:05 am

    @Philip Taylor
    This f***ing car won’t start!

  23. Andrew Usher said,

    November 17, 2022 @ 8:26 am

    Francois Lang:

    Exactly; that contains several instances of pointing out the difference between 'big' and 'little' in terms of order, and it's clear to me that it is 'big' and not 'little' that's the exception.

  24. Tucker said,

    November 19, 2022 @ 5:23 pm

    I watched Slumdog Millionaire again recently and I always found the game show host saying "Ricky Ponting, the Australian great cricketer" instead of "Ricky Ponting, the great Australian cricketer" jarring

  25. Philip Anderson said,

    November 19, 2022 @ 6:00 pm

    I wonder if that reflects Indian English, or the speaker was hinting that it was rare for a great cricketer to be Australian (rather than say Indian)?

  26. Michael Watts said,

    November 20, 2022 @ 7:31 am

    As for ‘dumb little’ vs. ‘little dumb’ , I think that the syllable constraint takes care of the lion’s share of the variance.

    This theory would seem to have a hard time explaining the phrase "little old lady". My instinct says that, in current use, the phrase does not describe a particular type of "old lady" (a little one); rather, it is itself a fixed expression in which all the words are basically parallel to each other.

  27. Michael Watts said,

    November 20, 2022 @ 7:38 am

    For little in particular, I've noted the fixed phrase "little old lady", but it seems worth observing that there are a few conventional collocations involving the word: you frequently hear "…cute little [X]", "…cool little [X]", and (maybe a bit less frequently) "…neat little [X]". In these cases, I don't think little has much in the way of semantic value; these compound descriptors aren't necessarily calling [X] small. As I think about it, they may be calling [X] metaphorically small, which little would not normally do.

  28. Benjamin Geer said,

    November 21, 2022 @ 8:02 am

    There's a lot of debate among linguists about how to explain adjective order restrictions. Here's a recent overview of these debates:

  29. J Lopez said,

    November 22, 2022 @ 4:33 am

    I think in English little+object often forms a semantic unit that functions as a diminutive. Come here, little boy. Say hello to my little friend. Little attaches itself to the noun in a way that big does not, and thus resists being broken apart. If you add another adjective, it will modify the entire phrase: (dumb (little dog)), (ungrateful (little child)). However, when there isn't a diminutive intent, then normal rules apply.

    On Google, "little bright star" (normal order since color follows size) and "bright little star" (diminutive order) both give about the same number of hits (304,000 vs 310,000). Whereas "little bright angel" (normal order) yields only 33 results to the 120,000 results of "bright little angel," presumably because angels are often described diminutively.

    Side note: James H. above says:

    That hat makes you look stupid, so it looks dumb on you. That makes it a dumb looking hat. Or just a dumb hat, as the "looking" doesn't seem to add anything to the meaning. You could say "stupid hat" just as easily.

    That's not quite the case. "Looking" does add meaning because it can clarify between a hat that looks dumb and a hat which is dumb for other reasons beside appearance. For example, if your hat is too light and flies off your head at the slightest draft, you might irritatedly call it a "dumb hat" even if it looks quite smart. Conversely you might have a hat about which you say, "I know this hat is a bit dumb-looking, but it really keeps me warm in sub-zero weather." In such a case, you'd be loath to call that hat just a "dumb hat," because it fulfills its purpose well.

RSS feed for comments on this post